Communication difficulties in ASD involve a speech reception deficit, whose biological causes are not yet identified. Neuropsychological studies reported dysfunctions from divergent perspectives, ranging from deviant behavior to different cortical activities and even the activation of different cortical areas compared to typically developing individuals. Among a number of hypotheses the communication deficit could, at least in part, result from atypical neuronal ensemble activity, as reflected by neural oscillations. In particular, atypical coupling across oscillation frequencies could disrupt the joint tracking and prediction of dynamic acoustic stimuli, both essential for speech comprehension. Computational models of speech perception have shown that coupled theta-gamma oscillations is instrumental to parsing and segmenting continuous speech into sub-lexical units (such as syllables), as it packages gamma activities into theta oscillation aligned to the amplitude envelope of the speech signal. Thus the brain can orchestrate more efficiently the bidirectional information flow, and coordinate internal predictive timing and knowledge deployment (beta activity). Disruption of synchronization across cortical circuits during speech processing is argued to provide a plausible neurophysiological explanation for deficits in speech processing. The nature of processing deficits underlying the ASD phenotype could hence be well accounted for by the perturbation of cortical synchronization. However, whether such oscillation anomalies can already be found in very young children with ASD, and with what specificity they relate to deficits in individual language reception capacity is unknown. In this thesis we explored neural activity of 64 very young children with and without ASD (mean age 3) using electroencephalography (EEG), while they were exposed to naturalistic continuous speech via an age-appropriate cartoon video. Such movie trailers embedded both visual and auditory stimuli. In the first study, we controlled the gaze pattern to reduce the cross-modality effect in the speech process in both groups. We performed EEG power analysis to extract oscillatory activities involved in speech processing and phase-amplitude coupling to detect the cross-frequency coordination in speech process. Speech envelope reconstruction from corresponding EEGs were measured via multiple linear regression to present neural entrainment. Furthermore, we established the specific relevance of oscillation anomalies with respect to speech development. EEG power typically associated with phrase-level chunking (delta, 1-3Hz), phonemic encoding (low-gamma, 25-35Hz) and top-down control (beta, 12-20Hz) was markedly reduced in ASD relative to typically developing (TD) children. We also found comparable theta activates across ASD and TD. Critically, children with ASD showed alterations of theta/gamma coupling and an atypical beta/gamma coupling. Speech neural-tracking by delta and theta (4-8Hz) oscillations was also weaker in ASD than TD children. Even though many oscillation features were atypical in our ASD sample, the beta/gamma coupling anomaly was the single best predictor of individual speech reception difficulties. Furthermore, this feature predicted only speech reception scores and none of the other cognitive functions.